Neural crest origin of sympathetic neurons at the dawn of vertebrates

Abstract

The neural crest is an embryonic stem cell population unique to vertebrates1 whose expansion and diversification are thought to have promoted vertebrate evolution by enabling emergence of novel cell types and structures like jaws and peripheral ganglia2. While basal vertebrates have sensory ganglia, convention has it that trunk sympathetic chain ganglia arose only in jawed vertebrates^3–8. In contrast, here we report the presence of trunk sympathetic neurons in the sea lamprey, Petromyzon marinus, an extant jawless vertebrate. These neurons arise from sympathoblasts near the dorsal aorta that undergo noradrenergic specification via a transcriptional program homologous to that described in gnathostomes. Lamprey sympathoblasts populate the extracardiac space and extend along the length of the trunk in bilateral streams, expressing the catecholamine biosynthetic pathway enzymes tyrosine hydroxylase and dopamine β-hydroxylase. CM-DiI lineage tracing analysis further confirmed that these cells derive from the trunk neural crest. RNA-seq of isolated ammocete trunk sympathoblasts revealed gene profiles characteristic of sympathetic neuron function. Our findings challenge prevailing dogma which posits that sympathetic ganglia are a gnathostome innovation, instead suggesting that a late-developing rudimentary sympathetic nervous system may have been characteristic of the earliest vertebrates.

The advent of the vertebrate lineage some 550 million years ago ushered in a remarkable expansion in the morphology and physiology of chordate body plans⁹. This expansion was enabled in large part by the emergence of the neural crest¹⁰, a transient population of migratory stem cells that contribute to numerous structures important for the vertebrate lifestyle including most of the craniofacial skeleton, outflow tract of the heart, peripheral sensory and enteric ganglia, among others¹. Though many neural crest derivatives were present in ancestral jawless vertebrates, others, in particular jaws and sympathetic ganglia, reportedly arose later in gnathostomes^3–8. Thus, extant jawless agnathans—lamprey and hagfish—are important models for understanding the emergence of complex vertebrate forms and traits¹¹.

The neurons of the sympathetic chain ganglia and the chromaffin cells of the adrenal system both arise from sympathoadrenal progenitors^12–16. These in turn originate from neural crest cells that emerge from the dorsal neural tube and migrate ventrally, coalescing around the dorsal aorta to forming sympathetic chain primordia. In the vicinity of the dorsal aorta, neural crest cells are exposed to high concentrations of BMPs, which in turn induce expression of transcription factors critical for establishing sympathoadrenal fate^17,18. Among those are Ascl1, Phox2b, and Hand2. Downstream of these core transcription factors lie the catecholamine biosynthetic pathway processing enzymes tyrosine hydroxylase (Th) and dopamine β-hydroxylase (Dbh) which are required for nor/epinephrine biosynthesis¹⁹, a defining characteristic of noradrenergic and adrenergic cells. A second wave of sympathoadrenal progenitors arise from late-arising neural crest-derived Schwann cell precursors that form fate-restricted chromaffin cell progenitors within the adrenal primordium, which lies in close association with sympathetic chain primordia^20,21.

The evolutionary origin of sympathoadrenal-lineage cells in vertebrates is not well understood. Agnathans were reported to lack peripheral sympathetic ganglia^3–8, though sparse clusters of chromaffin cells have been reported in both lamprey and hagfish^4,22,23. Nonetheless, these agnathan chromaffin cells differ from those of gnathostomes in several key aspects, including their lack of extrinsic innervation^24,25, which in gnathostomes is provided by sympathetic neurons. Furthermore, the transcriptional program conferring sympathoadrenal specification in gnathostomes was not detected during early neural crest development in lamprey⁵, suggesting that agnathans lack cell type(s) homologous to gnathostome sympathoadrenal-lineage cells. Thus, the embryonic origin and specification of agnathan adrenal chromaffin cells, and the subsequent expansion and diversification of the sympathoadrenal lineage in gnathostomes, remain unresolved.

Conserved sympathoadrenal specification

Despite the reported absence of sympathetic chain ganglia in jawless fish^3–8, the presence of sparsely distributed chromaffin cells in extant agnathans^4,22,23 is consistent with the possible presence of a potential sympathoadrenal progenitor. To further explore this possibility, we performed in situ hybridization chain reaction (HCR) to examine the spatiotemporal expression in embryonic lamprey of key transcription factors critical for establishing sympathoadrenal fate in gnathostomes. We identified robust expression of core transcription factors—Ascl1, Phox2, and Hand—in lamprey embryos as early as T25. Ascl1 was present in the anterior head region and at lower levels throughout the notochord, whereas Phox2 showed strong expression in the midbrain, hindbrain and surrounding the pharyngeal arches. Finally, Hand was largely localized to the heart region and the ventral pharyngeal mesenchyme (Extended Data Fig. 1). Notably, and consistent with prior reports⁵, these genes were not co-expressed in any population of cells at early stages.

Unexpectedly, analysis of later-stage embryos—T27 and older—revealed expanded expression of the core sympathoadrenal transcription factors. By T28, Ascl1, Phox2, and Hand were expressed in clusters of cells that spanned the heart and extended into the trunk (Fig. 1a–i). Importantly, the three transcription factors were widely co-expressed (Fig. 1j–n). These findings raise the intriguing possibility that the canonical sympathoadrenal transcriptional specification program identified in gnathostomes may be highly conserved, albeit deployed only at late embryonic stages in lamprey.

Figure 1: Co-expression of sympathoadrenal fate-specifying genes in late embryonic lamprey.

(a) HCR detection of Phox2 (teal), Ascl1 (white), and Hand (red) in lamprey at T28. DAPI is shown in blue. Red asterisk denotes the heart. (b-i) Co-localization of all three transcripts is seen in cells spanning the trunk dorsal to the yolk tube. Scale bars=100μm. (j-n) Transverse section of multiplex HCR in T28 lamprey as shown in (a). DAPI is shown in blue. Bilateral cells co-expressing Phox2 (teal), Ascl1 (white), and Hand (red) reside above the yolk tube, flanking the dorsal aorta. White asterisk denotes the dorsal aorta. Scale bars=50μm. n=15 embryos across five independent replicates.

Catecholaminergic cells in lamprey trunk

Cells of the sympathoadrenal lineage are molecularly and functionally defined by their biosynthesis of the catecholamines, norepinephrine and epinephrine. Accordingly, we observed co-expression of Th and Dbh genes—which encode two processing enzymes required for norepinephrine biosynthesis—with core transcription factors Ascl1, Phox2, and Hand by HCR (Extended Data Fig. 2). Th and Dbh were themselves co-expressed as early as T27 (Extended Data Fig. 3a–d), and we detected expression of the norepinephrine transporter Slc6a2 as early as T28 (data not shown). Cells co-expressing Th and Dbh were clustered in the ventral part of pharynx, around the heart and into the trunk at the level of the yolk tube, and by T28, spanned the length of the trunk in a bilateral chain-like arrangement (Fig. 2a–j). Rather than being assembled into ganglia, however, these cells appeared as single or small clusters of cells.

Figure 2: Expression of catecholamine biosynthetic enzymes by lamprey sympathoblasts.

(a-j) HCR detection of Th (teal) and Dbh (red) in lamprey at T28. Co-localization of the transcripts is seen in cells surrounding the heart (b-d) and spanning the initial length of the trunk in bilateral streams (e-j). Scale bars=100μm. n=15 embryos across five independent replicates. (k-m) Immunohistochemical detection of TH (teal) at T28 is consistent with HCR. HNK1 antibody staining of axons is shown in red. Scale bars=100μm. n=15 embryos across five independent replicates. (n, o) Immunohistochemical detection of TH (red) at (n) T27 and (o) ammocete stages. At both stages, TH⁺ cells are localized lateral to the dorsal aorta (denoted by white asterisks) and dorsal to the yolk tube or the intestinal epithelium. TH⁺ cells are highlighted in zoomed insets. DAPI is shown in white. Scale bars=50μm. n=12 embryos, 12 ammocetes across four independent replicates.

Immunohistochemical analysis of TH expression yielded consistent results (Fig. 2k–m), and further clarified the cells’ localization. Analysis in transverse sections at both T27 and ammocete stages at mid-trunk axial level confirmed TH⁺ cells were localized in close proximity to the dorsal aorta and outside the lumen of the intestine (Fig. 2n, o), consistent with their identify as sympathoblasts. While the heart-associated sympathoadrenal cells we observed are consistent with the chromaffin cells previously described in lamprey^22,23, the bilateral chains of cells flanking the dorsal aorta along the trunk are suggestive of peripheral sympathetic neurons, which have not previously been observed in basal vertebrates.

Late sympathetic maturation in lamprey

To further characterize these putative lamprey sympathetic neurons, we compared TH⁺ cell morphology across early and late developmental stages. This analysis uncovered subtle but significant changes in the appearance of cells in the trunk: from T27 to T30, TH⁺ cells became progressively more differentiated, showing increasingly refined and elongated primary processes (Fig. 3a–d; Extended Data Fig. 4). This morphological trajectory is suggestive of sympathetic trunk neuroblasts, which undergo extensive migration en route to their target sites within the trunk and adopt a more stereotyped neuronal morphology upon maturation^15,26. By ammocete stages, we detected the presence of the pan-neuronal marker Neurofilament-M in TH⁺ cells, confirming their identity as sympathetic neurons (Fig. 3e–k). Notably, expression of Neurofilament-M in TH⁺ cells was not detected at T30, the latest embryonic stage, nor in ammocetes prior to four months of age (Extended Data Fig. 5a–f). On the contrary, in ammocetes four months and older, we consistently observed expression of neuronal markers Neurofilament-M, HuC/D, and SCG10 (Extended Data Fig. 5g–o). Taken together these results suggest that, though lamprey sympathoblasts appear to give rise to sympathetic neurons, neurogenesis is not complete until later larval stages.

Figure 3: Late-developing sympathetic neurons in the lamprey trunk.

(a-d) Comparison of cell morphology based on immunohistochemistry for TH at T27 (a, b) and T30 (c, d). Scale bars=10μm. (e-k) TH (teal) and Neurofilament-M (Nf-M, red) co-expression is detected by immunohistochemistry in streams of neurons in ammocete trunks. Scale bars=50μm. n=15 embryos per stage and 15 ammocetes across five independent replicates.

Neural crest origin of sympathoblasts

In gnathostomes, the sympathoblasts that give rise to adrenal chromaffin cells and trunk sympathetic neurons are derived from neural crest cells^14,15 and the later-arising neural crest-derived Schwann cell precursors^20,27; however, the embryonic origin of these progenitors in agnathans is unknown. To address this important question, we performed fate mapping experiments using the lipophilic dye, CM-DiI. At T21, following neural tube cavitation, we injected CM-DiI into the lumen of the neural tube to label prospective neural crest cells (Fig. 4a, b). Injected embryos were then harvested at T27 and T30 to assess for colocalization with TH (Extended Data Table 1). Consistent with results in gnathostomes, we identified CM-DiI labeling in a subset of TH⁺ cells at both T27 (Fig. 4c–e; Extended Data Fig. 6a–c) and T30 (Extended Data Fig. 6d,e), demonstrating a conserved origin of lamprey sympathoblasts from trunk neural crest-derived cells.

Figure 4: Neural crest origin of lamprey sympathoblasts.

(a and b) Representative images of CM-DiI labeling of neural crest at T21, time of injection, in (a) whole mount with CM-DiI shown in gray, and in (b) transverse section with CM-DiI shown in red and DAPI shown in white. NT=neural tube. Scale bars=50μm. (c-e) DiI-lineage tracing of TH⁺ sympathoblasts at T27. CM-DiI-labeled neural crest (red) localizes to TH⁺ cells (teal), revealing a neural crest origin. DA=dorsal aorta, E=esophagus. Arrowheads indicate DiI-labeled cells in split-channel zoom. DAPI is shown in white. Scale bars=20μm. Experimental details are shown in Extended Data Table 1.

Lamprey sympathetic molecular identity

To gain insight into the molecular identity of lamprey sympathetic neurons, we harvested bilateral tracts of cells—presumed to be the sympathetic chains—from the trunks of lamprey ammocetes (Extended Data Fig. 7a–c). We further selected for TH⁺ and TH⁻ cells and subjected them to RNA sequencing using Fixed and Recovered Intact Single Cell RNA-seq (FRISCR)²⁸. Differential expression analysis comparing the transcripts of TH⁺ versus TH⁻ cells suggested significant commonality between sympathetic neurons of lamprey and other vertebrates (Fig. 5, Extended Data Figure 8, SI Table 1, and SI Table 2). Protein-protein interaction analysis revealed hubs including the TH-activating protein 14-3-3. PRKACA, which controls trafficking and localization of norepinephrine transporter was also indicated, further suggesting catecholaminergic identity. Synaptic protein hubs PSD-95 and SNCA, and developmental signaling hubs including cell adhesion-associated ITGB1 were also enriched (Fig. 5b).

Figure 5: Conserved molecular characteristics of lamprey sympathetic neurons.

(a) Volcano plot of differentially expressed genes identified between TH⁺ and TH⁻ cells harvested from ammocete trunks. Annotation is provided for the top upregulated transcripts. (b) GO analysis of Enrichr Protein-Protein Interaction (PPI) Hubs⁴⁹ for TH⁺ neurons. Analysis is performed using the top 500 significantly upregulated transcripts with an enrichment cutoff of 1.5 log₂(Fold Change). p-values are determined using Fisher’s exact test. (c) Heat map comparing expression of developmentally-regulated and canonical sympathetic (e.g., neuronal, transporter, and synaptic)²⁹ genes between TH⁺ and TH⁻ cells. Key indicates gene expression as log₁₀(TPM, transcripts per million). (d-o) HCR validation of mRNA-seq targets in ammocete trunks shown in transverse sections. Th (teal) colocalizes with Gphn (d-g), Nptn (h-k), and ATP6V1B2 (l-o), each shown in red. DAPI is shown in white. Scale bars=20μm. n=4 ammocetes across three independent replicates for each gene combination.

Given the function of both TH and DBH as oxygenase enzymes, and the role of iron as an essential cofactor for TH activity, that iron and oxygen showed significant enrichment in inferred metabolites analysis is consistent with the cells’ catecholaminergic identity (Extended Data Fig. 7d). Our analysis also implicated copper, an essential cofactor for DBH^29,30. Furthermore, as catecholamine biosynthesis within nor/adrenergic cells is confined to membranous vesicles, enrichment in terms including extracellular vesicle and membrane-bound vesicle lent further support (Extended Data Fig. 7e, f). Finally, comparison of developmentally-regulated (e.g., Sema3f and Sema3a) and canonical neuronal (e.g., Nefm and Nefh), transporter (e.g., Slc31a1 and Slc6a3) and synaptic (e.g., Snap25 and Rab3c) gene expression²⁹ between TH⁺ and TH⁻ cells revealed consistent upregulation of these genes in TH⁺ cells, altogether supporting the notion of basic homology amongst vertebrate sympathetic neurons (Fig. 5c).

To validate these results, we selected several targets that were significantly upregulated in TH⁺ neurons, as well as implicated in enriched processes or terms based on our GSEA results. Our analysis pointed to Gephyrin, a synaptic scaffolding protein encoded by the gene Gphn, which associates with TH-activating 14-3-3 as well as other postsynaptic density proteins. Another target suggested by our analysis was the adhesion molecule protein Neuroplastin, encoded by Nptn. Similar to Gephyrin, neuroplastins are enriched in neuronal synapses where they modulate synaptic function and plasticity³¹. Finally, we assessed ATP6V1B2 which encodes a subunit of the V-ATPase proton transporter and associates with 14-3-3 as well as synaptic proteins. Activity of V-ATPases generate electrochemical proton gradients necessary for vesicle and secretory granule loading in sympathetic neurons and chromaffin cells, respectively, and contribute to neurotransmitter release and exocytosis³². To analyze expression of these genes, we generated probes against Gphn, Nptn, and ATP6V1B2 for multiplex HCR in ammocetes. In transverse sections through ammocete trunks, we observed expression of each of the candidate genes in Th⁺ neurons (Fig. 5d–o). Collectively, these results serve to validate our RNA-seq analysis, and further suggest potential basic homology between the peripheral sympathetic chain neurons of lamprey and gnathostomes.

Discussion

Decades of comparative anatomy and observational studies have held that, in addition to prominently lacking jaws, agnathans fail to form sympathetic ganglia^3–8. Our present findings challenge this dogma, as we identify TH-expressing sympathetic neurons aligned in a chain-like manner in the trunk of the basal vertebrate lamprey. These neurons derive from trunk neural crest cells, and apparently are specified towards sympathoadrenal fates in response to a conserved transcriptional program. How then have trunk sympathetic neurons evaded detection through decades of prior studies? The sympathetic neurons of lamprey arise at much later developmental stages than in humans and well-studied amniote model organisms^26,33,34, such that direct comparative analyses would precede their specification. Moreover, compared with amniotes and even later-developing teleost fishes35, the number of lamprey sympathoblasts is limited; there are likely to be fewer sympathetic neurons present in larval and mature lamprey, complicating their histological detection. Whether late-developing trunk sympathetic neurons were a general feature of agnatha during early vertebrate evolution remains unknown, but our findings raise this intriguing possibility. Importantly, our results challenge the notion that the sympathetic trunk is an entirely gnathostome innovation.

While we found that lamprey sympathoblasts deploy the same transcriptional specification program described in gnathostomes, their sympathetic neurons differ from those of jawed vertebrates in several significant respects. First, the timing of sympathoblast specification and subsequent neurogenesis are significantly delayed in lamprey. We did not detect the presence of neuronal markers in TH⁺ cells until ammocete stages; even at T30, the latest embryonic stage, we observed no colocalization with mature neuronal markers. Second, the abundance of sympathetic neurons in post-embryonic stages critically differ in lamprey^26,33,34. We observed sympathoblasts, and subsequently sympathetic neurons spanning the length of the trunk in sparse bilateral chains, though at no point did we observe the formation of well-defined ganglia in either late-stage embryos or ammocetes. Interestingly, these key differences appear to be in keeping with broader observations of gradual sympathetic ganglia evolution in gnathostomes. Anatomical studies in cartilaginous fishes (Chondrichthyes), including dogfish³⁶ and the chimaera, spotted ratfish^37,38, revealed diffuse and poorly organized sympathetic chains of the trunk, as well as the absence of well-defined and interconnected sympathetic ganglia. By comparison, bony fishes (Osteichthyes) present considerably more elaborate sympathetic trunks³⁹, with teleost fishes’ most closely resembling those of tetrapods⁴⁰. Nonetheless, in the teleost zebrafish, trunk sympathetic neurons are initially sparse, and do not form appreciable ganglia until four weeks post-fertilization³⁵, considerably later than in amniote models. Altogether elaborated trunk sympathetic ganglia appear to have emerged gradually in gnathostomes, with Chondrichthyes boasting a modest sympathetic trunk that only became fully interconnected in Osteichthyes.

Beyond the apparent differences in timing of development and anatomy of the sympathetic chains, there may be differences in the molecular heterogeneity of sympathetic neurons between fish and amniotes. Detailed transcriptional profiling of thoracic sympathetic neurons in mouse has revealed the presence of five molecularly distinct noradrenergic sympathetic neuron subtypes (NA1-NA5), in addition to two cholinergic subtypes (ACh1 and ACh2)⁴¹. While these different populations have yet to be fully characterized, they are believed to imbue the sympathetic system with specific target-dependent output. For example, NA2 and NA5 populations selectively control nipple- and pilo-erector muscles, respectively. NA3, an abundant population characterized by NPY expression, is proposed to represent vasoconstrictor neurons, which regulate vasculature tone via innervation of blood vessels in smooth muscles⁴². In our analysis of lamprey sympathetic neurons, we detected Npy co-expression in Th⁺ neurons by HCR, suggestive of the NA3 population (Extended Data Fig. 9a–c); however, we found no evidence for the proposed NA2 and NA5 noradrenergic subtypes in lamprey (Extended Data Fig. 9d–k). Because nipple- and pilo-erection are mammalian-specific arousal responses, the NA2 and NA5 populations that regulate these behaviors may have only arisen in mammals. By contrast NA3 neurons, which are far more abundant and also appear to be present in teleost fishes, may represent an early noradrenergic population.

The acquisition of jaws imbued gnathostomes with increased predatory capacity^43,44, which in turn placed significant selection pressure on the evolution of defensive traits^45,46, such as an elaborated sympathetic trunk. The sympathoadrenal—so-called fight-or-flight—response is an ancient adaptation to environmental and predatory threats through which the body is mobilized for action⁴⁷. In this response, a perceived threat stimulates the release of epinephrine from adrenal chromaffin cells, which has the effect of mobilizing glucose from the liver, accelerating respiration, and diverting energy from digestion. Lamprey chromaffin cells, in addition to their restricted distribution in the heart^22,23, appear to be incapable of epinephrine biosynthesis owing to the absence of a phenylethanolamine N-methyltransferase (PNMT) ortholog⁴⁸. Nonetheless, norepinephrine—which acts on the same adrenergic receptors as epinephrine—is present in lamprey. This raises the possibility that a rudimentary sympathetic response mediated by norepinephrine may have been active in basal vertebrates. Gnathostomes would have then progressively expanded and elaborated upon this ancestral system with the addition of epinephrine (via the emergence of PNMT), expanded adrenal chromaffin cell distribution, and increased sympathetic neuron cell density and connectivity in the trunk.

METHODS

Animals

All animal experiments have been reviewed and approved by the Institutional Animal Care and Use Committee (IACUC) of Caltech, Protocol 1436. Adult sea lamprey (Petromyzon marinus) were obtained through the Great Lakes Fishery Commission. Adults were housed and spawned in standard conditions¹. Embryos were staged according to Tahara². Ammocetes were analyzed at 2.5 months, 4 months, or later (14–16cm of unknown age). At these stages lamprey cannot be accurately sexed. Sample size was subject to availability, and randomization was not required. Data for morphological analysis were blinded.

In situ Hybridization Chain Reaction (HCR)

Probe pairs were generated against Petromyzon marinus Ascl1 (XM_032949360.1), Phox2 (XM_032955429.1), Hand (XM_032956541.1), Th (XM_032949364.1), Dbh (XM_032976431.1), ATP6V1B2 (XM_032950546.1), Gphn (XM_032965246.1), Nptn (XM_032975984.1), Npy (XM_032959215.1), Ret (XM_032969454.1), Gfra (XM_032950358.1), and Slc6a2 (XM_032971063.1) genes. Labeling and detection procedures were performed as described by Molecular Technologies’ HCR v3.0 kit and reagents: http://www.moleculartechnologies.org/supp/HCRv3_protocol_zebrafish.pdf for whole mount and http://www.moleculartechnologies.org/supp/HCRv3_protocol_generic_slide.pdf for transverse sections³.

Whole-mount Immunohistochemistry

Lamprey embryos and ammocetes were fixed in 4% paraformaldehyde overnight at 4°C, then stored in 100% methanol at −20°C. Prior to staining, samples were rehydrated through graded methanol/PBS washes, ending in 100% PBS. Samples were then permeabilized with 10μg/ml Proteinase K for 20–90 minutes (time dependent on stage), followed by twenty minute post-fix in 4% paraformaldehyde. Later-stage samples were further permeabilized in buffer containing 33% DMSO in 3.3% PBS-Triton-X for three hours. Samples were then equilibrated to 0.1% PBX-Triton-X and blocked (10% normal goat serum in 0.1% PBX-Triton-X containing 0.5% BSA and 1% DMSO) at 4°C overnight. The following day, samples were incubated with primary antibodies diluted in blocking overnight at 4°C. Following washes in 0.1% PBS-Triton-X, samples were incubated with secondary antibodies diluted in blocking solution overnight at 4°C. Antibodies were used at dilutions as follows: rabbit anti-Tyrosine Hydroxylase (Milipore Cat#:AB152, 1:500), mouse anti-HNK1 (DSHB Cat#:3H5, 1:50), mouse anti-neurofilament (Invitrogen Cat#:13–0700, 1:200), and rabbit anti-Dopamine beta-hydroxylase (Immunostar Cat#:22806, 1:500). Alexa Fluor secondary antibodies were used at a concentration of 1:500.

Immunocytochemistry

Lamprey embryos and ammocetes were fixed in 4% paraformaldehyde at 4°C overnight, then washed in PBS before sinking in 30% sucrose at 4°C. Sunken samples were embedded in OCT and sectioned by cryostat at 20μm thickness. Prior to antibody labeling, sections underwent sodium citrate antigen unmasking according to DAKO target retrieval solution protocol (Cat#:S1699). Sections were then blocked (10% normal goat serum in 0.1% PBX-Triton-X containing 0.5% BSA and 1% DMSO) for one hour at room temperature, then incubated with primary antibodies diluted in blocking overnight at 4°C. Following washes in 0.1% PBS-Triton-X (PBS-TX), sections were incubated with secondary antibodies diluted in blocking solution overnight at 4°C. Sections were then washes in PBS-TX, counterstained with DAPI, and mounted for imaging. Antibodies were used at dilutions as follows: rabbit anti-Tyrosine Hydroxylase (Milipore Cat#:AB152, 1:500), mouse anti-neurofilament (Invitrogen Cat#:13–0700, 1:500), mouse anti-HuC/D (Invitrogen Cat#:A21271, 1:500), and mouse anti-SCG10 (DSHB Cat#:L5/1, 1:100). Alexa Fluor secondary antibodies were used at a concentration of 1:1000.

CM-DiI Labeling of Neural Crest

Lamprey embryos were injected with CM-DiI (ThermoFisher, Cat#:C7000) as previously described^4,5. Briefly, CM-DiI was diluted (1:5) in 10% sucrose and microinjected into the neural tube of T21 stage embryos. Injected embryos were screened to ensure adequate expression and appropriate targeting, then grown to T27 or T30 and fixed in 4% paraformaldehyde at 4°C overnight. Following fixation, embryos were prepared for embedding and sectioning as described above. Numerical description of CM-DiI analyses are provided in Extended Data Table 1.

Image Acquisition and Analysis

All images were acquired on a Zeiss LSM 980 inverted confocal microscope with Carl Zeiss Zen Blue software maintained by the Beckman Imaging Facility of Caltech. Minimal processing for brightness and contrast, and pseudo-coloring were performed in Adobe Photoshop (2023). Analysis of relative cell morphology was performed by measuring the ratio of primary neurite length to diameter of the soma. These measurements were made as described by Kim et al., 2015, with the neurite length encompassing the tip of the neurite to the soma boundary⁶. The soma diameter is measured as the length between the two most distant foci on its periphery. Statistical significance was determined by the Mann-Whitney U-test, which is appropriate for comparison between two independent groups in which the dependent variable is not well modeled by the Normal distribution. GraphPad Prism 9 was used for analysis and graph generation.

Cell Selection and cDNA Synthesis

Lamprey ammocetes (average length of 14–16cm) were sourced directly from the Great Lakes. Bilateral longitudinal fiber tracks dorsal to the intestines and ventral to the notochord were dissected out of the trunks, and these cell clusters were resuspended in pre-warmed Accumax (Innovative Cell Technologies). Samples were partially digested in a 37°C thermomixer shaking at 1400rpm for 16–20 minutes, followed by trituration with fire-polished glass Pasteur pipets to yield single-cell suspensions. Cells were fixed and stained according to Thomsen et al.⁷ using rabbit anti-TH (Milipore Cat#:AB152 at 1:1000) and Alexa Fluor 488, goat-anti-rabbit secondary (1:1000). TH⁺ cells and adjacent TH⁻ control cells were manually selected as described by Hempel et al.⁸, reverse crosslinked, and purified using the NucleoSpin RNA XS kit (Takara, Cat#:740902.10). cDNAs were synthesized using the SMARTseq v4 ultra low input kit (Takara, Cat#:634891), and purified using the AMPure XP PCR purification kit (Beckman Coulter, Cat#:A63880). Per replicate, ~300 cells were harvested from three ammocete trunks to yield three TH⁺ and three TH⁻ control samples for sequencing.

RNAseq and Bioinformatic Analyses

Purified cDNAs were transferred to the Caltech Millard and Muriel Jacobs Genetics and Genomics Laboratory where quantity and quality were assessed by Bioanalyzer (Agilent) prior to library construction using the NEBext Ultra II kit (NEB). Libraries were run on an Illumina HiSeq2500 high throughput sequencer at a depth of 30 million reads. Reads were mapped to the Stowers Institute SIMRBASE Petromyzon marinus reference genome, kPetMar1.0 (https://simrbase.stowers.org/sealamprey?pane=resource-1) using Kallisto⁹. Differential expression was determined using the DEseq2 package in R¹⁰. Gene ontology analyses were performed using Enrichr^11–13 with input of the top 500 most significantly upregulated genes with a log₂(Fold Change) enrichment cutoff of 1.5. Genes analyzed by heatmap are from terms GO:0061549 (sympathetic ganglion development), GO:0097492 (sympathetic neuron axon guidance), and GO:0097490 (sympathetic neuron projection extension), as well as canonical sympathetic neuronal, transporter, and synaptic genes previously experimentally validated¹⁴. For comparison between lamprey and mouse sympathetic neurons, previously reported mouse noradrenergic neuron data were used¹⁵. Analysis was performed using the 500 most abundantly expressed genes across all noradrenergic neuron subtypes.

Extended Data

Extended data figure 1: Expression of sympathoadrenal genes at early embryonic stages.

(a-c) HCR detection of Phox2 (teal), Hand (red), and Ascl1 (white) in lamprey at T25. (d) The three transcripts are not co-expressed. Scale bar=50μm. n=6 embryos across three independent replicates.

Extended data figure 2: Co-expression of sympathoadrenal fate-specifying genes with catecholamine biosynthetic enzymes in post-embryonic lamprey.

HCR detection of Th (teal) and Dbh (red) with (a-d) Ascl1, (e-h) Phox2, and (i-l) Hand in white at T30+ in whole mount. The core sympathoadrenal transcription factors are persistently co-expressed with catecholamine synthesis enzymes. Scale bars=20μm. n=9 embryos across three independent replicates.

Extended data figure 3: Early expression of catecholamine biosynthetic enzymes.

(a-d) HCR detection of Th (teal) and Dbh (red) in lamprey at T27. Colocalization of the transcripts is seen in cells surrounding the heart and spanning the initial segment of the trunk in bilateral streams. Scale bars=100μm. Red asterisk denotes the heart. n=6 embryos across three independent replicates.

Extended data Figure 4: Quantification of relative morphology of TH⁺ neuroblasts/neurons from T27 to T30.

Morphology of TH⁺ cells was compared from T27- and T30- staged embryos by plotting [neurite length/soma diameter]. This measure revealed increased neurite length in T30 as compared to T27, consistent with greater neuronal maturation. **** p<0.0001 by two-tailed Mann-Whitney U-test. A total of fifty cells from five embryos were measured for each developmental stage. Error bars indicate SEM.

Extended data figure 5: Late neurogenesis of lamprey sympathoblasts.

(a-i) Immunohistochemical detection of TH (teal) and Neurofilament-M (Nf-M, red) at T30 (a-c), 2.5 month ammocete (d-f), and 4 month ammocete (g-i) stages. Co-expression of TH and Nf-M is not observed in late embryonic or earlier ammocete stages. Consistent co-labeling is observed by four months of age. n=12 embryos, 12 ammocetes of each stage per marker across four independent replicates (j-o) Co-expression of TH with additional neuronal markers is observed in four month old ammocetes. HuC/D (j-l) and SCG10 (m-o) are shown. DAPI is shown in white. Scale bars=50μm. n=9 ammocetes of each stage per marker across three independent replicates.

Extended data figure 6: Lamprey sympathoblasts are neural crest-derived.

Lineage tracing of TH⁺ sympathoblasts. CM-DiI-labeled neural crest (red) localizes to TH⁺ cells (teal) at (a-c) T27 and (d, e) T30, indicating a neural crest origin. Arrowheads (b, c, and e) indicate CM-DiI-labeled cells. DAPI is shown in white. Scale bars=20μm. DA=dorsal aorta, E=esophagus, I=Intestine. Experimental details are shown in Extended Data Table 1.

Extended data figure 7: Analysis of gene expression in lamprey sympathetic neurons.

(a) Representative image of a wild-caught lamprey ammocete used for selection of trunk sympathetic neurons. Scale bar=1cm. (b and c) Immunohistochemical detection of TH (teal) in transverse sections through ammocete trunks. DAPI is shown in white. Yellow box indicates the site of the sympathetic chain neurons. SC=spinal cord, NC=notochord. Scale bars=250μm (a) and 10μm (b). n=4 ammocetes across three independent replicates. (d and e) GO analyses of Metabolites and Cell Compartment for TH⁺ neurons. Analyses are performed using the top 500 significantly upregulated transcripts with an enrichment cutoff of 1.5 log₂(Fold Change). p-values are determined using Fisher’s exact test. (f) Immunohistochemical detection of DBH (teal) at T28. Punctate expression is consistent with vesicular localization of DBH protein. Scale bar=20μm. n=6 embryos across three independent replicates.

Extended data figure 8: Comparison of gene expression between mouse and lamprey sympathetic neurons.

(a) GO analysis of Enrichr Protein-Protein Interaction (PPI) Hubs⁴⁹ for mouse sympathetic neurons previously reported⁴¹. Analysis is performed using the 500 most abundantly expressed genes across all noradrenergic neuron subtypes. p-values are determined using Fisher’s exact test. (b) Venn diagram indicating PPI Hub terms implicated in both mouse and lamprey sympathetic neurons.

Extended data figure 9: Analysis of noradrenergic subtypes in lamprey.

HCR detection of Th (teal) with (a-c) Npy (red), (d-g) Npy (red) and Ret (white), and (h-k) Gfra2 (red) and Ret (white) in whole mount ammocete trunks. Consistent co-expression of Th and Npy (a-c) suggests the presence of the NA3 noradrenergic subtype in lamprey. A lack of consistent co-expression of Th with Npy and Ret (d-g) and Gfra2 and Ret suggest the absence of noradrenergic subtypes NA2 and NA5, respectively. Scale bars=20μm. n=12 ammocetes across four independent replicates.

Extended data table 1: Number of sympathoblasts labeled by CM-Dil injection at T21.

Stage	Total no. embryos analyzed	Total no.CM-Dil+/TH+ cells
T27	5	12
T30	5	19

Data Availability

All raw RNA-seq data generated from this study are publicly available through NCBI’s GEO database under accession GSE246248. RNA-seq data used for comparison to mouse sympathetic neurons were published previously⁴¹ and are available under accession GSE78845. Source data for Extended Data Figure 4 are provided. Raw data can be made available upon request.